Optimal early detection and prevention strategies for breast cancer depend on our ability to accurately identify individuals with significantly increased risk for this disease. Unfortunately the ability of the medical community to predict risk of breast cancer accurately for individual women remains limited. Risk assessment for other cancers is enhanced significantly when the tissue at risk can be examined for premalignant features (models include cervix, colon, esophagus, etc). This suggests that a tissue-based strategy could enhance risk prediction for breast cancer. Of note, the diagnostic categories seen in breast biopsies with benign findings (so-called benign breast disease [BBD]) - non-proliferative, proliferative without atypia and atypical hyperplasia - are accepted by many to form a continuum in the progression toward breast cancer. This histological assessment can serve as a base to which other markers of risk can be added for the construction of tissue based risk prediction strategies for breast cancer. To this end, we have assembled a large cohort of women with BBD identified at Mayo Clinic. Over the period of this proposed grant, the cohort will be expanded to include ~ 13,000 women who had biopsies 1967-2001 and who will have developed an expected ~ 1200 breast cancers. For the women in this Mayo BBD Cohort, we have collected extensive histologic features along with clinical and epidemiologic risk factors. Original benign tissue will be available for all women included, and these BBD tissue blocks will be used in the study of select biomarkers. In our work to date, we have shown that robust discriminatory features are detectable within the benign tissue. These include the overall diagnostic category of BBD, the extent of the proliferative process (e.g. number of foci), the extent of involution of background terminal duct lobular units and the expression of COX-2, the first biomarker tested. This further motivates our hypothesis that a comprehensive model that incorporates histologic and molecular features from benign breast tissue can enhance the precision of breast cancer risk prediction. The proposed work encompasses three aims that will be performed in a nested case-control series within the cohort: 1) test the calibration and discriminatory accuracy of the three available risk prediction models that incorporate BBDrelated data (Gail, Tyrer-Cuzick, Colditz-Rosner); 2) build a tissue-based risk prediction model that incorporates histologic and molecular features, along with clinical and epidemiologic risk factors and 3) validate the best model from aim 2 in an independent set of cases and controls from Mayo and two external validation sets from BBD cohorts at Vanderbilt and Henry Ford Hospital. With the assessment of the performance of risk prediction tools in women with BBD, and the construction and validation of a tissue-based risk prediction model, we aim to improve the accuracy of breast cancer risk assessment. PUBLIC HEALTH RELEVANCE: Accurate prediction of who is at increased risk for breast cancer is essential to identify those women who would benefit most from heightened surveillance and risk reduction strategies. Women who have had breast biopsies with benign findings (approximately 1 million US women/year) are known to be at increased risk for a later breast cancer. In their tissue we have found strong correlates of risk of a later breast cancer. We will combine tissue-based histologic and biomarker information, with standard risk factor data, from a large cohort of women with benign breast biopsies to craft a comprehensive risk prediction model for breast cancer.